Big data in practice - what it's used for sector-by-sector
There are truly no bounds to the scope and breadth of what can be achieved with big data. Here, we detail some of the most prominent applications of big data within leading industries and – where necessary – by job function.
This is by no means an exhaustive list – and it never could be. However, we want to map as many as we can. If you have any other applications of big data that are not listed here, please email them to mark.young@bigdatainsightgroup.com and we’ll add them to list.
Click the industry sector or job function to be directed to that area or simply browse through the list.
General manufacturing
Utilities
Car making
Tech start ups/apps developers
Retail and marketing
Finance
Insurance
Policing
HR
Sport
Gambling
Healthcare and pharmaceutical
• Predictive maintenance scheduling. Leading manufacturers place sensors around their machines to build an understanding of their usual patterns of operation and then detect frequency changes which indicate that the machine will soon need maintenance. They can then book the machine in when convenient, rather than causing unplanned downtime (which is a big problem) or servicing the machine unnecessarily.
• Simulations. Manufacturers can take real data from their products on the market and then run simulations based on what would happen if they changed one particular component or design aspect. They can then find ways to make the product cheaper, more reliable or more environmentally friendly. The Formula 1 racing teams are particularly adept in this area, as are advanced aerospace companies.
• Expanded product design modelling. Similarly, with new big-data enabled computer aided design programmes, product designers can substitute components or materials from huge databases and then access in-depth information on how this affects the final product, including the ramifications on cost, production processes, environmental effects, legislative requirements, supply chain and so on. Adding cloud into the mix allows thousands of different variations of a product to be rendered in a fraction of the time that it previously took to render one.
• Asset monitoring. As with the machines in manufacturing plants, the utilities companies use big data to keep track on all of their assets spread across a country, continent or the globe. This enables them to fix any broken asset (such as a sewage cleansing plant, a leaking pipe or a gas pump), perform pre-emptive running maintenance or isolate areas in which repair actions have been ineffective. Real time analysis with simple dashboard visualisation can save them time, manpower and money.
• Fault logging and cost predictions. Car makers place hundreds of sensors on components around the car which constantly log data on performance and faults. All of this data can be used to reengineer designs for more efficient products and to predict what the strain of warranty repairs are likely to be on cost and man resource.
By taking performance data from all of their cars, car makers can see which geological and ecological factors influence the performance of the car and wear and tear of the parts. They can then design bespoke models to suit certain markets.
Tech start ups/apps developers
• Partnering for new revenue streams. Leading web and mobile developers are among the most innovative and prosperous users of data. A mobile app, for instance, generates reams of data which is recorded continually and can be analysed in real time. For instance, listening patterns on popular music streaming services can be analysed against location data to build an understanding of where sub cultures are arising.
Tech start ups use this information to improve their own service offerings but also to ‘partner’ with other organisations to create new revenue streams from insight. For instance, record companies can buy information about trends from song identification apps or ‘second screen’ social media collators in order to monitor the popularity of their acts and to see which geographical locations the artist is making an impact in. They can then advertise accordingly to make an intervention where needed.
• Mood mapping. Retailers use feeds from social networks to build an understanding of how their products and company reputation is seen among the public. With the constant streams of opinions from Facebook, Twitter, Google+ and the like, companies are able to cheaply and quickly gather large samples of customer opinion. This means they can beta test products or ideas for ventures and then amend them as per consumer sovereignty, or simply get a feel for how they as a company are perceived and make interventions accordingly. The ability with new big data tools to turn unstructured data of this kind into a structured form that can be analysed is one of the biggest forces behind the big data momentum.
• Near field communication. Forerunners in the retail industry are now beginning to use near field communication (NFC) in their billboard and pop up stand adverts which allows offers, coupons, video content and the like to be transmitted directly to mobile phones when an enbaled device comes within a few feet of the chip. In return, the advertiser receives information about how many people engaged with each particular advert in each particular location and how many of them then went on to engage further, either by downloading more content or following links to websites. Advertisers can also see where any offers they provided are then cashed in, giving them an understanding of which promotional buttons are pushed with consumers and also how far they are willing to travel to redeem said offers and which geographical areas are more likely to take them up.
• Ad retargeting. Online retailers can ‘recapture’ lost sales by retargeting the customer on other websites. When a prospective customer views a particular product on a website but doesn’t go through and make the purchase, a redesigned advert for the product (or one similar) is displayed on other websites that the target goes on to view. This redesign will include techniques to hook the customer that differ from the initial ones that failed to secure the sale, such as different point-of-sale information, pictures or aesthetics.
• Loyalty cards. Not all big data applications are new. One of the oldest and most prominent is still the most effective. The store cards that supermarkets offer are ostensibly marketed as loyalty cards which entice customers to continue shopping with one brand in order to build up discount points. But they are more valuable to the supermarket in illustrating customers’ buying behaviours and for marketing new lines. Where items are placed on shelves, their cost, the promotions and advertisings attached to them and many more things beside all come together to heavily influence sales. The patterns of buying behaviours mapped by store card allows the effects of subtle changes on each demographic of society to be understood and therefore manipulated.
• B2B supplier profiling. Finance professionals can use big data to check on the ‘health’ of their suppliers and business partners. They can monitor a variety of indicators including when creditors pay their bills and whether there is any change in the normal patterns, analysed against indicators such as share prices or particular market conditions for the industry sector. They can then use this information to identify which companies might be a credit risk.
• Fraud detection. Companies like Visa are using big data to create fraud detection models which can flag up potential fraudsters. If set patterns are identified (such as spending over £100 at a petrol station followed by taking £200 out of a cash machine after 2am) then the credit card company can immediately notify the customer involved and take the necessary action.
• Credit Scoring. Similarly, the credit scoring that banks perform on their customers takes into account many factors which change often and sometimes rapidly, including details of their financial assets, transactions, previous lending history and demographics. This is then used to decide whether the customer can access loans and bank accounts.
New short term lenders (‘pay day loan’ companies) analyse data in real time to approve loans for customers who’s credit ratings do not allow them to apply for regular borrowing lines.
• Premium costing. Like the banks, insurance companies use a glut of indicators in determining the cost of their policy premiums. This includes factors relating to the car (the age, performance, security, value etc), the person (age again, driving history, employment etc) and location (where the car is parked, where it is driven, crash rates in the area etc).
Leading car insurance companies are now beginning to offer ‘black box’ services. They install a device into a customer’s car which monitors multiple aspects of their driving, including where they drive, how long for, at what speeds, where they park and more. The company is then able to offer premiums which are costed up according to risk. The idea is that young drivers are able to get cheaper policies if they drive sensibly, rather than having to pay high premiums because they’re tarred with the same brush as their riskier peers. This system also enables the insurance company to feed back to the customer on how they are driving by text message or email, allowing them to make an intervention.
Furthermore, location data such as GPS is enabling insurance companies to pinpoint accident black spots for the first time, meaning they can avoid having to charge a higher premium to anyone in the surrounding area and they can feed this information to public authorities who can investigate the exact causes of the high concentration of incidents.
• Suspect tracking. By combining CCTV images, facial recognition software, travel trends and identifiers on travel cards, police forces can capture criminals by automatically linking people to their likely destinations on buses and metro systems. This allows police to catch those that they miss at the scene of the crime and also to control arrest statistics, meeting targets for arrests in one London borough, for instance, as needed.
Predicting the whereabouts of criminals can also be made more efficient by advanced analytics of a variety of factors including the time, weather and events that are happening.
• Identifying leavers. Identifying which employees were looking to leave a business used to be an inexact science. Now, HR managers can set up automatic alarms which are triggered when an employee reaches a certain index score, comprised of a mixture of certain activities like emailing out CVs, downloading large amounts of data from CRM systems, visiting certain websites and so forth.
• Talent spotting. The bestselling book-cum-movie Moneyball, starring Brad Pitt in the film adaptation, is all about big data. Specifically it is about how data is used to help make decisions in sport. Fiction it isn’t, either. The story documents the work of Billy Beane, general manager of the Oakland Athletics baseball team in the US, who put together a team of champions at a rock bottom price from a pool of apparent losers by examining their stats with advanced analytics. The method is said to have been repeated by UK football clubs including Liverpool, Newcastle and Arsenal.
End users can see hoards of stats on football players, including calculations of their contribution to a game in real time during matches, using a second screen application called Squawka.
• Odds calculator. Already one of the most mature users of data and advanced analytics, big data has allowed bookmakers to be even more accurate when they set their odds and also to open up extremely valuable new revenue streams. The companies are now able to have their odds automatically fluctuating by the second online, live during sporting events, based on what is happening. If a substitute enters the pitch during a game of football, he is added to the goal scorers odds list instantly; if there is a corner, the odds on ‘next goal is a header’ are either dramatically lowered or are temporarily suspended.
• Crowd sourcing. ‘Hack-a-thons’ and ‘code-a-thons’ – events where teams of data analysts and computer science whizz kids compete against one another to find a solution to a real world problem – are becoming prominent across all industries. However, it is the healthcare and pharmaceutical sectors which have blazed the trail. The benefits of this kind of crowd-sourcing are principally that you garner innovative, ‘outside of the box’ solutions to problems because the people working on them aren’t bound by the perspectives embedded within institutionalised employees. Secondly, two heads, as they say, are better than one. A hundred, it follows, are better still. But while employing one hundred data analysts becomes expensive, a top prize of a few thousand pounds should be enough to entice a sizable field in a competition. Cloud based delivery of data and parallel computing for analysing it has enabled this development.
• Mobile record retrieval. Elsewhere within healthcare, centralised systems are being developed which allow doctors to research protocols for the treatment of specific conditions on mobile devices, in order to help with patient care.
Got an application of big data that isn’t detailed above? Email mark.young@bigdatainsightgroup.com and we’ll add it to the list.
